PROJECT SUMMARY/ABSTRACT Using Socially Assistive Robot Assistants to Augment NeuroRehabilitation Exercise Therapy Robots have the potential to tremendously impact the standard of care and improve the lives of patients and the working situation of existing caregivers. Socially Assistive Robotics, or robots used to help people through social, rather than physical, interaction has the potential to benefit several communities. In particular, the repetitive nature of neurorehabilitation for Traumatic Brain Injury (TBI)/stroke provides an ideal application domain for the use of a robot in care settings. The long-term goal of this research is to develop an effective care agent for neurorehabilitation for patients with TBI/stroke. This research will greatly advance the field through the engineering development and the community engagement required to move this research from the lab to the clinic setting. Furthermore, the novel study of observer responses to agents with varying embodiment will shed light into the response of an in-person embodied agent is different from a phone or computer agent on a screen. The objective of this project is to develop a robot prototype system for use in care environments through a user-centered design and evaluation process. The central hypothesis is that embodied robots can provide effective assistance to a human caregiver during TBI/post-stroke neurorehabilitation and that a system can be developed for a robot to be effectively integrated into a caregiver's current workflow. The rationale for this research is that neurorehabilitation is time-intensive and has better results with supervision. There is a workforce shortage of people qualified to provide this supervision and that current technology can enable robots to be able to supervise a patient for small windows of time and address the care shortage. This will help improve patient compliance which may lead to better health outcomes. The central goal of this project will be furthered through three specific aims: (1) the user-centered design and in-lab validation of a robot system designed for use with patients with TBI to support their caregivers; (2) a long-term evaluation of the systems performance and its effect on the workflow in a clinic setting that extends the research out of the lab and into care environments for multi-session evaluation; and (3) behavioral and neural experiments that compare an observer's response to an embodied agent to an on-screen 2D or 3D agent. Studies of Human-Robot Interaction primarily consists of short-term interaction scenarios. Results from this work is usually bounded by the time spent interacting. This project will address the needs of the Socially Assistive Robotics community by approaching generalizable problems related to long-term Human-Robot Interaction and applying them to studies of neurorehabilitation for TBI/stroke. The contribution of the proposed research is that it will lead to significant advancements in long-term Human-Robot Interaction enabling a richer impact on patients' lives. This contribution is significant because it will demonstrate how robots can be employed effectively in a care setting given the needs of both the patient and caregiver.